I am looking for some help with the object detection. I have object detection 10 lines of python script up and running with my USB camera, i noticed it is using gstreamer to process the video stream and display it. I would like to modify the gstreamer process parameters to point the stream to my localhost lamp stack using tcpserversink. I have the correct command line parameters to send video from my USB cam on the nano to my localhost:8080 and in properly displays within the video element of my webpage, however i would like to implement the Object detection into the stream sent to my localhost LAMP stack. I was thinking that since the 10 lines of python code example uses C++ bindings to interpret the python to c++, that if i found the line of code that initiallizes the gstreamer i could modify it to use tcpserversink and point it to my localhost:8080 stack.
1.) Where is the init line that calls gstreamer for output
2.) Is this the best way of accomplishing my goal?
2.a.)If not, does anyone have any suggestions?
Hi,
We suggest use DeepStream SDK. you can install the package through SDKManager. There are C code samples in
/opt/nvidia/deepstream/
The samples are demonstraed with ResNet 10 and Yolo models.
For python samples, please check